This article is rated Start-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||||||
|
Daily pageviews of this article
A graph should have been displayed here but graphs are temporarily disabled. Until they are enabled again, visit the interactive graph at pageviews.wmcloud.org |
This article was the subject of a Wiki Education Foundation-supported course assignment, between 5 September 2018 and 12 December 2018. Further details are available on the course page. Student editor(s): Chenruduan.
Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 02:37, 17 January 2022 (UTC)
A variety of mergers has been proposed. Please add to the discussion at Talk:Linear regression.
"The variance is also not restricted to being normal." --145.107.9.241 (talk) 12:13, 21 April 2008 (UTC)
Notation is inconsistent: transpose is indicated sometimes by T, sometimes by '. —Preceding unsigned comment added by 216.195.149.49 (talk) 02:55, 6 July 2008 (UTC)
This article is misleading and redundant. It is misleading because, for no apparent reason, it largely consists of a derivation of the ML estimator under normality. But it is a common fallacy that the errors in linear regression must be normal. Although the article does not explicitly state the errors must be normally distributed, the reader could draw that impression. It also contains a number of other problems: it fails to specify what makes the model "linear," it says the assumption X is of full column rank means the "parameters" are not linearly dependent when ought to say "covariates," it claims the errors must have a mean of zero and spherical distributions neither of which is generally required, it claims the X's are fixed (also an unnecessary assumption), and it fails to mention the ML estimator of \sigma^2 is biased. Most of these problems could be fixed, but why? The article is redundant: there's nothing here shouldn't instead be discussed in the articles on least squares, linear least squares, or linear regression (possibly others too: the number of redundant articles on basic regression models is staggering). I suggest deleting this article and redirecting to linear regression. Sked123 (talk) 21:46, 11 August 2008 (UTC)
which are uncorrelated random variables each with expected value 0 and variance σ2
A year passed since the complaints were made that this article only presented ML theory (the ML method wagging the estimation dog, since problems of ML estimation are minimized).
I incorporated the LS estimation material in another subsection, so the article is more useful. The LS estimation of coefficients article isn't easy to find, imho. Kiefer.Wolfowitz (talk) 21:16, 27 June 2009 (UTC)
(June 2009) There are already too many articles on regression analysis containing too much overlap. More importantly, there needs to be room for an article that outlines all the uses of "linear model" in statistics, as this is not confined to regression. Logically this should go under "linear model", while stuff about specically about regression should go in regression-specific articles. Melcombe (talk) 09:25, 29 June 2009 (UTC)
Are phi the non-linear (or linear) functions of the independent variables X? ie phi1 is the function (linear or non-linear) of X1? Granzer92 (talk) 14:19, 17 February 2024 (UTC)